106 X1V.-The Oxides of the Elements uizcl the Law. By R. M. DEELEP, F.G.S. LAST session I had the honour to communicate to the Chemical Society a paper describing “ A New Diagram and Periodic Table of the Elements’’ (Trans., 1893, 63, 852). Since this was written, my attention has been called to R very interesting paper, by Professor K. W. Zenger, of Prague, published in the Sitzu?,gsberichte konig. Bohm. Ges. Wissenschaft. (1881, p. 408). I n this communication the author points out that the products obtained by multiplying the specific heats by the densities of the elements have values which are very similar for many elements having analogous pro- perties. To show this, he arranges the elements in I1 groups, the first group containing bromine and iodine, and the eleventh manganese, iron, cobalt, and nickel.Of course the result is only partially satis- factory, for we now know that elements belonging to the same classes have volume heats which, on the whole, decrease with increasing atomic weight, end that even series elements must be separated from odd series elements, although their volnme heats may be nearly iden- tical. I now purpose laying before the Society the outcome of some further work undertaken with a view to ascertain t o what extent a Consideration of the oxides supports or modifies previous conclusions ; for although in my previous paper some reliance was placed on the oxide-forming power of the elements, the diagrams were almost wholly based upon other periodic phenomena. Of course it is im- probable that any one periodic table can express the whole truth, for elements closely resembling each other in some respects may differ widely i n others.Still there does seem to be a fundamental periodic basis, in accordance with which the elements may be classified, b u t several other derived forms may be drawn t o illustrate the periodicity of the several physical or chemical properties. ~ o l u m e heats and volume atoms) mere plotted as ordinates upon a diagram, the abscissce of which were atomic weights. So regular were the curves and lines obtained by joining the summits of the ordinates, that it was possible to make use of the diagram for the purpose of separating the elements into classes and series, and, therefore, for constructing a periodic table based upon direct numvicrtl values. Lothar Meyer (Xodern Theories O J Chemistry, English Translation, tj 63) very clearly pointed out how such a diagram may be inter- I n my previous paper, certain valuesOXIDES OF THE ELEMENTS AND THE PERIODIC LAW.107 preted. He found that, as a general rule, elements having similar properties occupied similar positions on the several curves, and also that each wave of the diagram constituted a complete period. So useful has this method of interpreting the atomic diagram proved to be, that we may, I think, regard with some confidence the periodic variations of the ordinates with increasing atomic weight as valuable and trustworthy aids in investigations relating to the Periodic Law. Unfortunately our knowledge of the relative densities and specific heats of the elements, and even oE their compounds, is by no means complete.In the case of many elements, no determinations have been made, whilst other elements are known to exist in more than one physical condition, and to have two or more specific heats or densities. The fact that there is sometimes more than one value for the same element does not, however, seriously interfere with the symmetry of the volume heat diagram. The diagram Lothar Nejer had pi-eviously drawn had atomic volumes plotted upon it. He also interpreted the curves obtained in accordance with the periodic table drawn up by Mendelkeff. Being, however, led by the volume heat diagram to somewhat modify this table, especially as regards the grouping of the elements of low atomic weight, I have attempted to further elucidate the matter by plotting upon diagrams some other values Calculated from the relative densities of the oxides and the atomic weights of the elements. Although Brauner and Watts (Ber., 14, 48) have already proved that the niolecular volumes of the oxides do show periodic varia: tions, there do not appear to be any methods other than such as may be regarded as purely empirical, of diagrammatically illustrating the periodicity of the oxide-forming power of the elements and the periodic changes which the relative densities of the oxides undergo.One such method I have followed in constructing Diagrams 1: 2, and 3, (p. 115) which show some interesting results. -- Relative densities of oxides Atomic weights of elements- On these diagrams the ordinates = If t,he combination of the elements with oxygen produced com- pounds having densities which were the mean of their constituent atoms, ordinates calculated in this manner, and plotted upon a, diagram, mould give results similar to those of a volume atoms diagram.But the densities of the oxides do not follow such a law, and it is on this account, that the ordinates calculated in the above manner are interesting. Still, there is frequently a close agreement between the periodic variations in the densities of the elements and of their most stable cxides. In my previous paper, (Zoc. cit., p. 864) the elements were arranged in duplicate in one table, the odd and even series, 1, 2, 3, &c., giving108 DEELET: THE OXIDES OF THE ELEMEKTS what may be called the fundamental basis of the Periodic Law, and the duplicates l a , 2a, 3u, &c., actual or possible departures from it.The fundamental basis is shown in Table I (p. lll), and a special form in Table II (p. 112). In the third series, there are ten substances, all of which are regarded as elements. If there are really 10 elements in this series, there should also be 10 in the first, fifth, &c. ; but rather than adopt this view, I would suggest that either one of the substances in Series 3 is not an element, or if it be an element, its atgmic weight is not correctly given. Although Xi and Co have been placed in the same group, it is qnite possible that they are both true elements. Omitting the " typical element " fluorine, Mendelheff, in his table, makes man- ganese the only member of the left-hand column of his seventh group, spaces being left for the undiscovered elements required to complete the table as far as t'his group is concerned. It would, therefore, per- haps, be better to regard either manganese or iron as the redundant element.Diagrams 1, 2, and 3, although they do not furnish any evidence bearing on this particular point, show much the same general features as does the volume heat diagram, and in some respects confirm, or lather give additional support to, the fundamental periodic basis shown in Table I. The vertical scale of Diagram 1 Bas been made smaller than that of the other two diagrams, so as to admit of the ordinate of hydrogen being plotted upon it. Although I have drawn in curves and lines to render clear the general nature of the variation in the magnitude of the ordinates with increasing atomic weight, it must not be supposed that the variations from these cnrves or lines are necessarily due to illcorrect determinations of the densities, $c.Indeed there seems to be evidence that the lines and curves should have been drawn in duplicate, one set cut'ting the ordinates of the perissads, and the other those of the artiads. For instance, the perissads boron, nitrogen, aluminium, and phosphorus all stand above the curves, whilst the artiads carbon, silicon, and sulphur fall below. In Table 111 (p. 113) are given, A, the elements; B, the atomic weights of the elements ; C, the densities of the elements ; D, the oxides; E, the densities of the oxides; F, the volume atoms of the elements ; and G, the laat column, E t B.Diagram 1 has plotted upon it the values in column G correspond- ing to those elements haTing atomic meights less than that of potas- sium. The ordinate plotted for hydrogen is calculated from the density of ice, whilst for lithium it is calculated from the density of Li20, and so on. As on the volume heat diagram, the ordinate for lithium is shorter than that of hydrogen. The sniall numerals on theR.H . v E €LEY. -Tourn. C'hem. SOC. 226.13'94 DIAGRAM I II ATOMIC WEICHT- 4 0 5 0 6 0 7 0 8 0 - K C A s c TI V MN FE co NI c u CR Z N C A C E A s S E BR I SPECIFIC GRAVITY OF OXIDE ATOMIC WEIGHT OF ELEMENT .=(.n : 0 d 1 100 i10 120 130 % _ , Re Y SR Z R N B Mo R u Rti P o AC C O I N SN SS T K I SPECIFIC GRAVITY OF OXIDE ATOMIC WElCHT OF ELEMENT e = IAND THE PERIODIC LAW.109 diagrams give the valencies of the oxides plotted. As a general rule, these are the oxides of which the relative densities are well known. Many non-con secu tive oxides, the relative densities of which are known, have been omitted. In Table I (p. Ill), showing the probable fundamental periodic basis, a space occurs for an element having an atomic weight greater than that of hydrogen, and less than that of lithium. Also, owing to the low density of amorphous carbon, and the anomalous properties of the “ typical elements,” it mas suggested in my previous paper that the volume heats of the elements of the first series, for some reason or other, did not occupy positions on a line such as that of Series 3.Sodium and magnesium were also regarded as having departed from their positions on the curve of Series 2. These views receive additional support from the diagram of the oxides, for, omit- ting hydrogen, the ordinates plotted on Diagram 1 fall on two short lines and two curves, the lines and curves following each other alter- nately. On the first line, me have Li,O and BeO, and on the second line, Na,O and MgO. On the first curve fa11 Be20,, COO, and N,O,; whilst on the second curve there are A1303, SiOz, P205, and SO,. The series from lithium to fluorine is, therefore, exactly similar to that from sodium to chlorine, and they should be placed in the same order on the derived Table 11, although the fundamental order is as shown on Table I.However, on the derived Table 11, which has been reproduced from my previous paper, the typical elements have not been tabulated exxtly in this manner, the properties of the pure elements seeming to demand a rather different arrangement. To agree with the diagram of oxides, the following arrangement (p. 110) would be better. I may here call attention to the fact that the space nieasnred on the atomic scale between hydrogen and potassium is shorter than that between potassium and rubidium, and also than that between rubidium and caesinm. No doubt this is in some way connected with the anomalous group properties of the “ typical elements.” Iudeed it might be urged that on this account the elements of the first and second series should not be classed at all with those of the third, fourth, fifth, and sixth.At present our periodic arrangement should not be regarded as more than tentative; hut the accumulation of more data will doubtless eventually enable the true significance of u-hat has already been called the Periodic Lam to be realisecl. Turning now to Diagram 2, on which are plotted the ordinates for the oxides of the elements between chlorine and rubidium, we have, on the first line, K20, CaO, Sc203, and Ti02. But here a change takes place in the value of the ordinates, for V,05 and CrO, have considerably shorter ordinates thaE those of the four preceding oxides. Here Li and Be are in Series 1’. 1 2110 DEELEY: THE O D E S OF THE Series. L, ELEMENTSAND THE PERIODIC LAW. 111 Series of elements.7 -- 7- 3 .a. 4 4 k? 3 ? 4 CL.112 DEELES: THE OXIDES OF THE ELEMENTS 1 e % - G Fi -~ 3 n Y d' Series of elements. 1 L- r- c a. 4 tt I u c ' 3 c c c 1 z i5 .n . + k a. H n 3 r-4 k t- I_ n n 1 I I HAND THE PERIODIC LAW . 113 TABLE IIr . A . -. Elements . Hjdrogen ..... Lithium ...... Boron ....... Carbon ....... Sodium ....... Magnesium .... Aluminium .... Silicon ........ Phosphorus .... Beryllium ..... Nitrogen ...... .......... Sulphur ...... Potassium ..... Calcium ....... Scandium ..... Titanium ...... Vanadicm .... Chromium .... Manganese .... Iron .......... Cobalt ........ Nickel ....... . . . . . . . . .......... .......... Copper ........ .......... Zin :: .......... Germanium ... Arsenic ....... Selenium ...... Strontium .....Yttrium ...... Zirconium ..... Niobium ...... Molybdenum . . Ruthenium .... Silver ......... Cadmium ..... Indium ....... Tin .......... Antimou? ...... Tellurium . . . . / lodine ......... ............ . . . . . . . . . ........... . . . . . . . . B . .. Atomic weights . .- 1 -0 7 *02 9 -1 11 -0 12 -0 14 -03 2.3 04 24 -3 27 *01 29 3 3 31"03 32 -06 39 -14 40 -08 44 *1 48 -13 51"4 52 . 3 55 *0 56 *02 58 *i 58):6 63':1 65 -3 7, 7i13 79):0 75 *09 87 -5 89 -0 90 -0 94-0 95 -7 101 -65 107 -93 112'11 113 -7 119 -1 1.20 *3 12: -0 126 *83 ............. C . I D . Densities . 1 Oxides . I 0 *62 0 *59 1 -85 2 -6t! 1 *i 0 -985 1 '743 2 '583 2 -48 2 22 2 -07 0 -865 1 -37 . 9 ) . . . 5 *8i 7 '00 7 -39 8 -00 8.9 9 -0 8 -95 7 '15 7, 7 , .. i f 4 7 4'7 4 ' 2 2 -54 4 -15 7 -06 8 .6 12 9 6 10 -57 i'65 7 -42 7 -29 6 -7 6)'23 4 -95 .. - 72 H,O LisO B e 0 BsO, C 0 - N. Oj XgO 810. P.O. so. K20 CaO Sc. 0. R' a. 0 Al. 0. .. T10. .. v. 0 5 Cr03 3fIl.O. Fe.0. coo Co. 0. N i 0 pu'i203 CU. 0 zdb .. GeO. A+Oj Y Y SeO. SrO ZrO. Nb2O5 RuOs y 2 0 3 MOO. Ag. 0 CCiO 1.1.0. SnO. Sb. Oj Tdb. 1.0. .. E . .. Densities ... 0.917 2 -108 3 -086 1 7 9 1.457 1 -64 2 -805 3 -64 3 *94 2 -65 2 -22 2 -38 1 -936 2 -636 3 -25 3 -8 4 -25 4 -56 3 -35 2 -8 4 -75 5 '3 5 -6 4.8 5 -6 4 '8 6 -13 5 -75 5 '5 5 '73 4 '7 3 . '7 4-2 3'954 4 -5.1 5 -03 5 '85 4 '46 3 -92 7 '4 7 5 2 8 -2 8 -11 7 -18 6 - 7 3 i 8 6 -52 5 -1 4 -25 5 -02 F . Volume atoms of 3lenients . 0 *6149 0 -0806 0 -2029 0 -2429 0 '14 16 0 -0419 0 *Oil6 0 -0955 0 *0875 0 -0714 0 -0645 0 -022L 0 *0390 .- -.- . . . .I 0 -1141 0 -1 168 0 -1342 0 -1484 0.1515 0 *1534 0 -1367 0 -1094 0 -0i55 0 -3684 0 -0531 0 -029 0 *0461 0.0751 0 *0897 0 -1217 0 * 0987 0 -0771 0 -0651 0 * O f i l l 0 -0556 0 -049i 0 -0393 . . . . . . . . . G . ..- EiB . ..- 0 *9170 0 -2994 0 -3392 0 -1621 0 -1214 0 -1 168 0 -1217 0 -1498 0 -1458 0 -0935 0 -0.783 0 -0767 0 -0603 0 -0678 0-08 LO 0 -0561 0 -0883 0 -0947 0 -0651 0 4535 0 -0b63 0 -0946 0 -0954 0 -0817 0 *0955 0 -0819 0 -0966 0 * o m 0 -0842 0 -0877 0 -0650 0 -0492 0 -0559 0 -0500 0 -0515 0 -0565 0 -0650 0 ~04.74 0 -0409 0 -0728 0 50696 0 -0759 0 -05 23 0 -0631 0 -0562 0 9514 0 -05.21 0 -0408 0 -0335 0 -0395114 OXIDES OF THE ELEMENTS AND THE PERIODIC: LAW. , , ...... XTra-nium.. .... A. -- 2id S O Elements. ~~ Barium. ...... Lanthanum.. .. Cerium .......Tsntalurn ..... Tungsten.. .... Mercury ...... Lead ......... , . . . . . . . . ,, .... . , ....... , , ..... B. -- Atomic weights. 137 -0 13!! ‘6 -- lib: -3 lii *o , I 182 -5 200 *2 206 -93 TABLE 111-continued. C. -- Densities. D. Oxides. BaO La,03 3 , C& wb3 9 , Ta,O, HgO Bi203 Th02 P b 0 2 U S E. --- Densities. 4 -0 5 . 7 2 5 -9s 6 53 6 -0 6 -93 i -35 8 -01 7.32 11 -34 8 -89 8 -08 9 -S6 5 -26 10 *2 F. Volume atoms of elements. 0 -0291 0 *044S 0 -0456 9 -0574 ir -1048 0 9709 0 ‘0542 0.0469 0 -0476 0 9779 --- --- - - - - - 0 *0291 0.0417 0 -0428 0 -04‘7 1 0 -0427 0 -04993 @ *om2 0 ’0438 0 -0397 0 -0566 0 -04.29 0 -0388 0 -0424 0 ‘0438 0 ‘02 19 On Diagram 3, this is repented, fol- we have (Rb ?), SrO, YzOs, and ZrO, on a line, and then a decrease in the values of the ordinates of Nb,O, and Moo,.Tantalum and tungsten also show a similar change. In all three series the change first shows itself in elements of the fifth class of Table I. Owing to the high density ascribed to tungsten, the volume heat and volume atom of that element did not agree, OIL the volume heat diagram, with those of osmium and tantalam. From Table 111, it mill be seen that the density of the oxide is not abnoi.ma1. Where, on the diagrams, the ordinates are denoted by circles, the ordinates are not consecutive ones, and where halfmoons are given, the values have not been determined. By placing another element to fill in the gap between hydrogen and lithium, boron becomes the fifth element of ihe first series; indeed it becomes the analogue of vanadium. niobium, and tantalum. The same decreasein the value of the ordinate then occurs in all four cases (with coosecutive oxides) to the fifth element of the odd series, and although the chen;ical properties, and therefore the class oxides, of the elements of the first series are not in agreement with those of the third, fifth, and ninth, the fact that this change of density is common to all the series appears to give considerable support to the f andamental periodic basis indicated in Table I. OF course this classification somewhat alters the basis upon which MendeGeff predicted some new elements, for lie makes scandiumAVAILABLE “MINERAL” PLANT FOOD IN SOILS. 115 ekaboron, because it forms an oxide, SGO,, and becaase it occurs in the Same group as boron in his table. Fundamentally scandium is probably ekalithinm, and vanadium probably ekaboron. Similarly copper is ekasodinm, and zinc is ekamagnesinm. In column F, Table III, I bave given the volume atoms, so that they may be compared with the values in Column G. In the even series, there is, on the whole, a fairly concurrent change in the two values.